Rendement de l'éducation et choix des instruments. Application sur données vietnamiennes

This paper focuses on instrument choice while consistently estimating the returns to education in Vietnam. Using VLSS data, we explore different sets of exogenous instruments that rely on demand and supply side sources of variation in schooling as well as the matrix of instruments proposed by Hausman and Taylor [1981]. As in several studies, we find that IV estimates of the returns to education are substantially higher than the corresponding ols estimate. We show how the Hausman-Taylor matrix of instruments, when combined with other instruments, may be a useful way of consistently estimating an average return to education rather than a local average treatment effect (Angrist [1994]).